ZARZA We are Zarza, the prestigious firm behind major projects in information technology.
Listen "LLM Benchmarking and Evaluation"
Episode Synopsis
Analysis of Large Language Model (LLM) evaluation, detailing its foundational principles, diverse methodologies (including automated, human-in-the-loop, and LLM-as-a-judge approaches), and core quantitative metrics. It further critically examines the landscape and inherent limitations of LLM benchmarks and offers a detailed analytical review and comparative performance overview of leading open-weight models from various developers, categorizing them by architectural philosophy and specialization
More episodes of the podcast AI Intuition
Agent Builder by Docker
06/09/2025
AI Startup Failure Analysis
03/09/2025
AI Security - Model Denial of Service
02/09/2025
AI Security - Training Data Attacks
02/09/2025
AI Security - Insecure Output Handling
02/09/2025
AI Security - Prompt Injection
02/09/2025
Supervised Fine-Tuning on OpenAI Models
31/08/2025
In God we trust